Changelog

A new public, bilingual (DE/EN) catalog at usecases.serviceplan-agents.com walks you through real marketing and sales use cases our coworkers already handle — organized by department and pain point, not by feature list.

Pick your area, browse matching cases, and open any one of them to see the full signal → insight → action flow: what triggers the work, which coworker picks it up, what the deliverable looks like, and where a human stays in the loop. Every case links back to which of Hannah, Alex, or Elena owns it, so it's easy to see who to bring in for a similar job.

Our Trust Center has a new home and a refresh: security, privacy, and compliance information now lives at its own dedicated address, redesigned to match the rest of our public pages.

The content itself has been brought up to date — current sub-processor list (including where Mistral AI and our generative-media provider fit in, disclosed plainly rather than glossed over), current ISO 27001 and SOC 2 roadmap status, and refreshed answers across Overview, Controls, Resources, Subprocessors, and FAQ.

We'd rather state clearly where we're still building operational track record than imply certification we don't hold yet — that's the standard we're holding this page to going forward.

You can now check platform health at any time on our new status page, covering both the Coworker platform and the underlying Sokosumi platform:

• Coworker Console & API
• Coworker Email
• Coworker Chat & WhatsApp
• Coworker Task Processing
• Sokosumi Service
• Sokosumi API

Each component shows 90 days of uptime history, and past incidents are listed with what happened and how long they lasted — no more guessing whether something's on our end or yours.

Alongside Claude, the platform now runs select workloads on Mistral AI — a European model provider we've been rolling out gradually across chat routing, background task execution (like email triage and contact enrichment), quality-control judges, and audio transcription.

Voice messages are now transcribed with Mistral's Voxtral model instead of OpenAI's Whisper, and document processing — including OCR on scanned or image-based files — has been unified onto Mistral as well, giving us a consistent, cost-efficient pipeline for turning documents into usable text.

Nothing changes in how you interact with Hannah, Alex, Elena, Maya, or Jamal — this is a behind-the-scenes efficiency and resilience improvement, not a new feature to turn on.